DE102006057726A1 - Method for measuring electromagnetic radiation in aerospace instruments - Google Patents

Abstract

The invention relates to a method for measuring electromagnetic radiation in aerospace instruments, in particular for observing the earth or other celestial bodies. The object of finding a new way to measure electromagnetic radiation that allows both the detection of high expected signal levels without saturation and low signal levels with sufficiently high signal-to-noise ratio (SNR), without requiring gain switching, is achieved according to the invention in that parameters of the signal processing chains (1) of the detectors (11) are set in a targeted manner so that at least one detector (11a) detects the dynamic range (25) of the maximum expected signal level (S <SUB> MAX </ SUB>) of the electromagnetic radiation (2) and at least one further detector (11b) are adapted to a limited dynamic range (26) for small and medium signal levels to obtain a higher signal-to-noise ratio (SNR).

Description

The The invention relates to a method for measuring electromagnetic Radiation in aerospace instruments, in particular to observe the earth or other celestial bodies.

Instruments for measuring electromagnetic radiation have for each sensor element on a so-called signal channel, which is usually - as out 1 apparent - as a signal processing chain 1 from a detector 11 , an integrator 12 , an amplifier 13 and an analog-to-digital converter (ADC) 14 let represent.

Due to the impinging flux of electromagnetic radiation 2 is in the detector 11 generates an analog electrical signal that is summed up and amplified and finally converted into a digital output signal.

Each component of the signal processing chain 1 knows a certain signal level, from which it goes into saturation. Since above this maximum signal level, the signal is no longer evaluable, this provides a saturation limit 24 (see. 2 ) In the design of the signal processing chain 1 As a rule, it is important to ensure that the specific saturation signal level is not reached on any of the aforementioned components. Conversely, the choice of certain components defines the saturation limit 24 the evaluable with the instrument flow of electromagnetic radiation 2 ,

Has the incident radiation 2 a greatly varying signal level, the evaluation is further complicated by the fact that the signal at the input of the ADC 14 a noise component is superimposed. The noise is due to internal electronic noise processes 22 all upstream components of the signal processing chain 1 as well as external noise 23 (ie external noise) caused. All noise terms, in front of the amplifier 13 are particularly critical because they are reinforced with.

When Measure of the quality of the signals is the signal-to-noise ratio (SNR) defined.

Special applications of a measuring instrument make different demands on the signal processing chain to be designed 1 , Based on the application, an area of the signal level to be detected becomes dynamic range 25 - defined, which the instrument must be able to evaluate. If the intended application requires the evaluation of a large maximum signal level S _MAX , then the evaluability of small signals automatically deteriorates, since the SNR is too low there.

For certain application, therefore, in addition to the saturation limit 24 of the dynamic range 25 (maximum recordable signal S _MAX ) often with respect to the noise processes 22 and 23 minimal signal level 27 for which the SNR still allows a clear distinction of the useful signal from the noise component. This limit is specified as SNR _MIN and sets the threshold 28 for the SNR (abscissa parallel dashed line in 2 ) in this 2 are two curves K ₁ and K ₂ for two different sized signal processing chains in an SNR signal diagram 1 shown with different gain factors. With amplification factor, the following is the overall effect of the radiation conversion in the detector 11 to the output of the amplifier 13 designated.

Usually, the minimum required SNR is above the signal level - as in 2 for two different signal processing chains 1 shown - defined using only one value SNR _MIN . It can be seen that in the case of the solid curve K _1, although the entire dynamic range is observable, the signal quality in the lower range does not meet the predetermined minimum requirements for the SNR.

If you increase the amplification factor within the signal processing chain 1 , the dashed curve K _{2 is formed} , in the lower signal range of the required minimum signal level 27 can be detected by the SNR threshold 28 achieved in this point. However, the curve K ₂ already runs from the middle signal levels in saturation (saturation limit 24 ) and can be the upper part of the required dynamic range 25 do not cover.

Around this fundamental Solve a problem, are present in the art two approaches known:

1. Adjust the gain of the Signal processing chain by means of estimation based on previous knowledge of the expected signal.

One such method is e.g. in the instrument IASI of the European Weather Satellite METOP implemented. Here is the integration time of the entire Selection step divided so that in a first, very short integration interval, the signal level of the incident radiation only roughly determined and then with the same signal processing chain a second longer Integration with adjusted gain factor is performed. The loss of integration time is a disproportionate profit an adjustment of the amplification factor across from. The disadvantage is the realization of the different amplification factors by electronic switches, which in turn generate additional noise.

2. TDI (Time Delay Integration) method.

The so-called TDI method works without prefetching and subsequent gain adjustments. The TDI method is eg of EL Dereniak and GD Boreman: Serial Scan with Time Delay and Integration, in: Infrared Detector Systems, John Wiley & Sons, 1996, pp. 491-493 (ISBN 0-471-12209-2) described in detail. The object to be observed is recorded several times in succession and the measured values of the individual measurements are averaged. In scanning systems, this multiple scanning is realized via a plurality of sensor elements (usually arranged in a row) and one and the same object are recorded in close succession (the usual principle of satellite-based image recording systems). The amplification factors are the same for all sensor elements and adapted to the maximum signal level to be covered. Random differences in the sensitivities of the sensor elements are usually corrected electronically. The SNR improves when using the TDI method solely due to the averaging of the pixel signals of the uniformly sensitive sensor elements, but does not allow adaptation to varying signal levels.

In order to be able to cover a large dynamic range, in the above known methods, the amplification factor of the sensors is to be selected as a function of the maximum input signal to be detected. For this, however, it must be set at fluctuating signal levels to the maximum expected signal, whereby the SNR in the middle and lower part of the dynamic range is too low (see for illustrative curve K ₁ in 2 ). If the SNR is to be improved for small signal levels, as described above, either the sensitivity of the detector or the amplification of the signal processing chain must be switched or a TDI method must be used.

Of the Invention is based on the object, a new way for the measurement of electromagnetic radiation in instruments of the air and space exploration to find Earth or other celestial bodies, both the detection of high expected signal levels without saturation as well as low signal level with sufficiently high signal-to-noise ratio (SNR) allowed without switching the gain is required.

According to the invention Task in a method for measuring electromagnetic radiation in Instruments of aerospace for observation of the earth or other celestial bodies with multiple detectors, each detector with integrator, amplifier and Analog-to-Digital Converter (ADC) forms a signal processing chain, solved by that parameters of the signal processing chains of the detectors so specifically be set to at least one detector on a dynamic range for one maximum expected signal level of the electromagnetic radiation is set and at least one other detector on one restricted Dynamic range for small and medium signal level is set to in this limited dynamic range a higher one Signal-to-noise ratio (SNR) to obtain.

Advantageous be for the measurement exclusively Measurements of detectors allowed in their signal processing chain a parameter combination was set with which the current Signal level of the incident radiation below a saturation limit recorded dynamic range of the signal processing chain and detectors whose signal processing chain exceeded the saturation limit was excluded from the measurement. Preferably of several for the measurement uses approved detectors that signal in the signal processing chain set a parameter combination is where the highest SNR was achieved.

Of several approved for the measurement detectors that signal is selected appropriately, in the signal processing chain a parameter combination is set, in which a required SNR (SNR _MIN ) was exceeded by means of a shortest possible integration time to a Objektabbil To achieve an improved modulation transfer function (MTF).

The the parameter combinations associated with the detectors in the signal processing chains be for The application advantageously adjusted by targeted that they chosen appropriately, but to be permanently installed. Furthermore, it is expedient that from the ones assigned to the detectors in the signal processing chains Parameter combinations of at least one of the parameters at least a signal processing chain is set dynamically.

In an advantageous embodiment become the for the measurement operated detectors operated in a TDI mode and by means of a weighted summation to a measurement signal summarized, the weighting based on the in the signal processing chains set parameter takes place.

there can associated with the detectors in the signal processing chains Parameter combinations for the application in turn freely chosen but fixed or at least of the parameter combinations one of the parameters of at least one signal processing chain dynamically be set.

Advantageous can also individual detectors with fixed parameters of the signal processing chains and other detectors with dynamically adjusted parameters become.

For the setting different parameters, it is expedient that at least one the detectors are in their dimensions from the remaining detectors different.

In another embodiment In accordance with the invention it is advantageous that the detectors are in a sub-pixel sampling regime operated by shifts by fractions of the middle Pixel distance reaches a resolution increase becomes.

Further it is possible, to set different signal chain parameters the different Pre-allocate different filter elements detectors.

For the application In a TDI mode, the weighted summation is preferably implemented executed in analog or digital hardware. But there is also the Possibility, to implement the weighted summation implemented in software.

For special Sophisticated applications, it is possible that at one required high SNR above the SNR threshold in the associated signal processing chain the original Integration time to a shortened Integration time, for the SNR is still above the SNR threshold, to thereby read the next associated detector to achieve an increase in the MTF of the object image.

With the solution according to the invention it is possible a measurement of electromagnetic radiation in instruments of the air and space travel to observe the Earth or other celestial bodies, which both the detection of high signal levels without saturation as well as low signal level with sufficiently high signal-to-noise ratio (SNR) allowed without switching the gain is required.

The Invention will be explained below with reference to exemplary embodiments. The drawings show:

1 : a schematic view of a signal processing chain 1 from a detector 11 up to an A / D converter (ADC),

2 FIG. 2: a schematic representation of the SNR versus the signal level of the incident radiation for two signal processing chains with different amplification, FIG.

3 : Examples of possible arrangements of detectors 11 with respect to the direction of movement of the scanning instrument and representation of the respective scanning direction (solid arrow) and possible TDI direction (dashed arrow),

4 : A representation of a signal-to-noise behavior according to the method of the invention with differently set signal processing chains of associated detectors 11 , whereby the readability of large signal levels in the required high dynamic range 25 is possible and at the same time for small signal levels in a limited dynamic range 26 higher SNR can be achieved,

5 : a possibility to set different signal chain parameters by different sized active areas of the detectors 11a and 11b .

6 Three different possibilities, each with two groups of differently set signal chain parameters to combine the inventive method with a conventional Subpixelabtastverfahren or with the classical TDI method;

7 Two schematic representations of a scan of a target object 3 with relative movement of the detector 11 with a) a long integration time to summate the largest possible radiant energy and b) a shorter integration time to improve the MTF if a sufficiently high SNR is present;

8th : a signal chain parameter curve composed of signals from detectors 11 with differently set signal chain parameters, for a combination of the invention with a TDI method according to embodiment 3);

9 a hardware implementation according to embodiment 1) of the invention;

10 a hardware implementation according to embodiment variant 3) of the invention;

11 a hardware implementation according to embodiment 4) of the invention;

12 a software implementation according to embodiment variant 3) of the invention.

In 1 is a common signal processing chain 1 for a radiation sensor, which is present in the same way for each sensor element, if it is a multi-element receiver (line or matrix receiver). A sensor element which is characterized by a defined closed radiation-sensitive area will be referred to hereinafter as a detector 11 designated. It represents the beginning of a signal processing chain 1 in the further an integrator 12 , an amplifier 13 and an analog-to-digital converter (ADC) 14 having.

Due to the impinging flux of electromagnetic radiation 2 is in the detector 11 An analog electrical signal is generated in the integrator 12 added up and in the amplifier 13 is reinforced to finally in the ADC 14 in a digital output signal 21 to be changed.

Each of these components of the signal processing chain 1 has a certain saturation limit 24 of the signal level, from which the strictly monotonous signal waveform goes into saturation. From this saturation limit 22 the signal is no longer evaluable. The device with the lowest saturation limit 24 the usable signal level thus defines the detectable dynamic range 25 respectively. 26 the signal processing chain 1 for the incident radiation 2 ,

As the signal at the input of the ADC 14 is superimposed on a noise that is due to internal electronic noise processes 22 all upstream components of the signal processing chain 1 as well as external noise 23 (external interference signals), a signal-to-noise ratio (SNR) must be exceeded, so that the actual useful signal is detected as a measuring signal, which is defined as follows:

Based on the application is therefore in the signal processing chain 1 an area of the signal level - dynamic range 25 - to define, which the instrument must be able to evaluate. Furthermore, by the application within this dynamic range 25 usually a minimum requirement for the smallest signal level to be detected 27 , expressed as a threshold 28 SNR, a so-called SNR _MIN . These relationships are in 2 shown.

To control the final signal level (output signal 21 of the ADC 14 ) and thus the SNR can within the signal processing chain 1 except the (mostly via integrator 12 and amplifiers 13 Gain) alternatively other sizes can be changed.

• – spektrale Filter mit unterschiedlicher Transmission, wobei die Filter dabei in den Strahlengang vor die Detektoren 11 eingefügt oder alternativ oder auch zusätzlich direkt auf die Detektoren 11 aufgebracht werden;
• – die Empfindlichkeit der einzelnen Detektoren 11 (Einstellung oder Steuerung der Signalpegel einzelner Sensorelemente eines Arrays);
• – die Variation der Integrationszeit (z.B. am Integrator 12); oder
• – die Größe der aktiven (strahlungsempfindlichen) Fläche der Detektoren 11, wobei die aktive Fläche der Detektoren 11 durch die einmalig vorgewählte Kantenlänge (z.B. in Scanrichtung des Instrumentes verkleinert) oder – alternativ oder zusätzlich – durch teilweise Abschattung/Abdeckung der maximal hardwaremäßig verfügbaren aktiven Fläche variiert werden kann.

A radiation-proportional output signal 21 can also be set by:

• - Spectral filters with different transmission, the filter thereby in the beam path in front of the detectors 11 inserted or alternatively or additionally directly to the detectors 11 be applied;
• - the sensitivity of the individual detectors 11 (Adjustment or control of the signal levels of individual sensor elements of an array);
• - the variation of the integration time (eg at the integrator 12 ); or
• - The size of the active (radiation-sensitive) surface of the detectors 11 , where the active area of the detectors 11 can be varied by the only preselected edge length (eg reduced in the scan direction of the instrument) or - alternatively or additionally - by partial shading / coverage of the maximum hardware available active surface can be varied.

Below are the above, the output signal level 21 influencing measuring parameters (filter transmission, detector sensitivity, active detector surface, integration time, amplification factor) under the term signal chain parameters summarized.

The use of detectors 11 (also in the form of individual sensor elements of an array) with different sensitivity, different filters and different sized active surfaces usually has static character. That is, these variables are not or only with difficulty changeable during operation.

In contrast, the individual detectors 11 associated electronic quantities, integration time and gain factor, ie the dynamically variable parameters in the associated signal processing chain 1 , in principle, while the company is still free to set.

Basis of the invention are two or more detectors 11 , in succession, a target object 3 to capture. Possible changes of the observed target object 3 are comparatively small and negligible during this period. The detectors 11 can be arranged both in rows / rows or matrices.

3 shows examples of possible arrangements of detectors 1.1 a (not shown) radiation-measuring instrument for aerospace, which can absorb any radiation, preferably optical or radar radiation. In addition to the direction of movement of the instrument, the respective scanning direction (solid arrow) and the respective TDI direction (dashed arrow) are shown. Are the detectors 11 arranged in matrices, in principle, a two-dimensional TDI process can also be carried out.

Furthermore, the detectors must 11 not necessarily complete, as in 3 shown, be positioned. Rather, in practice a technologically or systemic distance between the individual detectors 11 to be available. This is permissible as long as the changes in the target object to be observed are related to the respective application 3 slow compared to the total duration of multiple sampling by the individual detectors 11 stay. In extreme cases, this condition can still be found in a combination of detectors 11 different instruments, which in turn are the same target object 3 to be staggered in time, to be fulfilled.

• 1) In einer ersten Ausführung der Erfindung wird mit Detektoren 11a, 11b (und ggf. weitere) zu unterschiedlichen aufeinanderfolgenden Zeiten das gleiche Zielobjekt 3 beobachtet. Diese Detektoren 11a, 11b usw. haben unterschiedliche Signalkettenparameter. Jede den Detektoren 11a, 11b ff. zugehörige Signalverarbeitungskette 1 liefert ein Ausgangssignal 21. Für die Auswertung des Messsignals wird das Ausgangssignal 21 desjenigen Detektors 11a, 11b ff. verwendet, dessen Signalkettenparameterkombination das größte Messsignal liefert und das gerade noch nicht in Sättigung ist, d.h. dessen Signalpegel unterhalb der Sättigungsgrenze 24 seines Dynamikbereichs 25 bzw. 26 liegt. Damit wird unter Ausnutzung eines eingeschränkten Dynamikbereiches 26 bestimmter Detektoren 11 das erreichbare SNR für kleine Signalpegel verbessert, ohne dass das Risiko besteht, plötzliche hohe Signalpegel nicht mehr messen zu können.

The parameters of the detectors 11 associated signal processing chains 1 are in all of the subsequent embodiments of the invention within the described rows, lines or matrices of detectors 11 not constant, ie there is at least one detector 11a which differs from the others in at least one signal chain parameter.

• 1) In a first embodiment of the invention is provided with detectors 11a . 11b (and possibly more) at different consecutive times the same target object 3 observed. These detectors 11a . 11b etc. have different signal chain parameters. Each of the detectors 11a . 11b ff. associated signal processing chain 1 provides an output signal 21 , For the evaluation of the measuring signal, the output signal 21 of that detector 11a . 11b ff. whose signal chain parameter combination supplies the largest measurement signal and which is just not yet saturated, ie its signal level below the saturation limit 24 its dynamic range 25 respectively. 26 lies. This is taking advantage of a limited dynamic range 26 certain detectors 11 Improves the achievable SNR for small signal levels without the risk of being unable to measure sudden high signal levels.

The principle of the procedure is in a hardware implementation in 9 shown. The detectors 11 are assumed here as pixels 1 to n of a sensor array, of which at least pixel 1 has a dynamic range 25 corresponding 4 for a maximum signal level S _{MAX to be} detected. For all subsequent pixels 2 to n are graduated limited dynamic ranges 26 (eg according to SNR curves K ₃ and K ₄ in 4 ), so that for their signals in saturation comparators 41 ₁ to 41 _n-1 It is checked whether a saturation (exceeding the saturation limit 24 ) is present. If the latter is the case, the signal will not be sent to the respective SNR comparators 42 ₁ to 42 _n-1 forwarded, each comparing whether one from a previous detector 11 (Pixel with a lower pixel number) of existing signal level is exceeded. These are the SNR comparators 42 ₁ to 42 _n-1 arranged in a pipeline structure, the speed of their processing (processing clock) exactly the movement of the individual forwarded signals of the associated detectors 11 about one and the same target object 3 must be adapted. This boundary condition is in 9 as supplied pixel clock r of the sensor array on which the detectors 11 are integrated, drawn. As a result, the signal of the detector 11 which has achieved the largest SNR (without its saturation limit 24 To exceed).

in the Difference to item 1 of the prior art must in this basic variant the invention available standing integration time can not be split. Thus, often too measuring small amounts of energy can be better used. Farther have to different gain factors not be switched so that the associated noise is eliminated. in the Difference from the prior art point 2 is for SNR improvement no TDI calculation necessary.

• 2) In einer zweiten Ausführung der Erfindung wird der maximale Signalpegel mit Hilfe eines oder mehrerer Detektoren 11 (i.d.R. des ersten oder der ersten Sensorelemente eines Sensorarrays) der oben beschriebenen Linien, Reihen oder Matrizen abgeschätzt oder ermittelt. Zu diesem Zweck werden die Signalkettenparameter dieses Detektors 11a bzw. dieser Detektoren 11a-11c so gewählt, dass der gesamte Dynamikbereich 25 abgedeckt wird. Die Signalkettenparameter der nachfolgenden Detektoren 11 werden anhand dieser Messung dynamisch so angepasst, dass das SNR für den nun eingeschränkten Dynamikbereich 26, der jedoch hinreichend bekannt ist, verbessert wird. Wegen der eventuell sehr ungenauen Messung/Abschätzung des aktuell benötigten eingeschränkten Dynamikbereiches 26 mit dem den gesamten Dynamikbereich 25 abdeckenden Detektor 11a ist die Verwendung eines oder auch mehrerer nachgeordneter Detektoren 11 sinnvoll. Wird nur ein nachgeschalteter Detektor 11b verwendet, so müssen dessen Signalkettenparameter so eingestellt werden, dass der abgeschätzte Signalbereich (eingeschränkter Dynamikbereich 26) mit Sicherheit nicht die durch die Signalkettenparameter eingestellte Sättigungsgrenze 24 überschreitet.

The signal chain parameters of the individual detectors 11 are basically freely combined adjustable to a desired signal gain in the output signal 21 of the ADC 14 to obtain. For example, for this method instead of signal processing chains 1 with detectors 11 same integration time and different gain factors and signal processing chains 1 with the same gain factor for detectors 11 be used with different integration time. Likewise, the output signals 21 also due to different sensitivity of the detectors 11 (eg pre-ordered filters or apertures) can be set.

• 2) In a second embodiment of the invention, the maximum signal level is determined by means of one or more detectors 11 (usually the first or the first sensor elements of a sensor array) of the above-described lines, rows or matrices estimated or determined. For this purpose, the signal chain parameters of this detector 11a or these detectors 11a - 11c so chosen that the entire dynamic range 25 is covered. The signal chain parameters of the following detectors 11 are dynamically adjusted based on this measurement so that the SNR for the now limited dynamic range 26 which, however, is sufficiently known to be improved. Because of the possibly very inaccurate measurement / estimation of the currently required limited dynamic range 26 with the the entire dynamic range 25 covering detector 11a is the use of one or more downstream detectors 11 meaningful. Will only be a downstream detector 11b used, its signal chain parameters must be set so that the estimated signal range (limited dynamic range 26 ) certainly not the saturation limit set by the signal chain parameters 24 exceeds.

Become several downstream detectors 11 their signal chain parameters can be adjusted to cover the blur range of the estimate of the expected signal level. In this case, the evaluation of the downstream detectors takes place 11 according to the above embodiment 1).

in the Difference to point 1 of the prior art also needs here the available standing integration time not to be split. Thus, can an incidental (low) amount of energy can be better utilized.

• 3) Eine dritte Ausführung der Erfindung basiert auf den in 3 gezeigten Sensorarrays in Form von Reihen/Zeilen oder Matrizen, in denen die Sensorelemente die einzelnen Detektoren 11a, 11b, 11c usw. bilden und unterschiedliche fest eingestellte Signalkettenparameter aufweisen. Die Signalkettenparameter können dabei für jeden einzelnen Detektor 11a, 11b, 11c, ... unterschiedlich sein oder aber es existieren Gruppen von Detektoren 11 (z.B. 11a-11c, 11d-11f usw.) mit in sich gleichen Signalkettenparameter, wobei sich die Gruppen untereinander jedoch in mindestens einem Signalkettenparameter unterscheiden.

In contrast to the prior art item 2, a significant SNR improvement takes place without TDI calculation.

• 3) A third embodiment of the invention is based on the in 3 shown sensor arrays in the form of rows / rows or matrices, in which the sensor elements, the individual detectors 11a . 11b . 11c etc. and have different fixed signal chain parameters. The signal chain parameters can be used for each individual detector 11a . 11b . 11c , ... are different or there are groups of detectors 11 (eg 11a - 11c . 11d - 11f etc.) with the same signal chain parameters, but the groups differ from one another in at least one signal chain parameter.

Using the readings of all non-saturated detectors 11 In addition, a measured value calculation is carried out according to the known TDI method. However, in comparison with the second point of the prior art, the SNR is significantly increased in the low signal range, without in the case of a large dynamic range 25 to be limited in the choice of signal chain parameters. The noise reduction of the measurement results is achieved only in that only all non-saturated by the signal level to be measured detectors 11 contribute to the determination of the measured value.

Eventually saturated detectors 11 will not be considered. Thus, the SNR is less improved at high signal levels than would be the case with classic TDI. In the lower and middle restricted dynamic range 26 On the other hand, there is an improvement that goes far beyond the values of classic TDI methods. Here is the. Averaging over detectors 11 with different signal chain parameters of the signal processing chain 1 with the additional use of weight matrices possible.

Example:

• – Setze die Anzahl der Detektoren 11 als einen rückwärts laufenden Zählindex m = n;
• – Vergleiche das Signal S_i (i = 1 ... n) der Signalverarbeitungskette 1 jedes Detektors 11 mit deren Sättigungsgrenze 24 (S_MAX),
• – Wenn S_i = S_MAX, setze die Signale S_i aller gesättigten Detektoren 11 auf Null und reduziere den Zählindex m um die Anzahl der zu Null gesetzten Signale S_i;
• – Bilde das Summensignal S_SUM der verbliebenen Signale S_i
  

The total signal of a target object 3 (only in 7 shown) can with the help of n detectors 11 from their known gain factors G ₁ ... G _n and the signal levels S ₁ ... S _n (output signals 21 at the exit of the ADC 14 ) are calculated using the following procedure:

• - Set the number of detectors 11 as a backward counting index m = n;
• Compare the signal S _i (i = 1... N) of the signal processing chain 1 each detector 11 with their saturation limit 24 (S _MAX ),
• - If S _i = S _MAX , set the signals S _{i of} all saturated detectors 11 to zero and reduce the count index m by the number of zero set signals S _i ;
• Form the sum signal S _{SUM of} the remaining signals S _i
  

There the (uncorrelated) noise comes from the root of the sum of the squares the individual measurements results increased itself the SNR accordingly.

Now only the parameter combination of the signal processing chain has to be done 1 a TDI stage (a TDI row), but at least that of a detector 11 , the entire dynamic range 24 cover. All other detectors 11 can increase the SNR in the region of low signal level so that a required minimum SNR (SNR threshold 28 ) can be achieved. Assign the detectors 11b . 11c with limited dynamic range 26 mutually different signal chain parameters, the improvement of the SNR is made above the signal level in the lower and middle part of the entire dynamic range 25 graded at intervals.

• 4) In einer vierten Ausführung der Erfindung wird der maximale Signalpegel wiederum mit Hilfe eines oder mehrerer Detektoren 11 (i.d.R. des ersten oder der ersten Sensorelemente der oben beschriebenen Linien/Reihen oder Matrizen) abgeschätzt bzw. ermittelt. Zu diesem Zweck werden die Signalkettenparameter dieses Detektors 11a oder dieser Detektoren 11a-11c so gewählt, dass der gesamte Dynamikbereich 25 abgedeckt wird. Die Signalkettenparameter der nachfolgenden Detektoren 11b, 11c bzw. 11d ff. werden anhand dieser Messung dynamisch so angepasst, dass das SNR für den nun eingeschränkten jedoch hinreichend bekannten Dynamikbereich 26 optimiert verbessert wird.

Such a construction of the invention is shown in the embodiment 1. In contrast to point 1 of the prior art, in this case the available integration time is not split and thus the amount of energy used is better utilized. Unlike the prior art item 2, the TDI calculation is based on different signal chain parameters.

• 4) In a fourth embodiment of the invention, the maximum signal level is again using one or more detectors 11 (usually the first or the first sensor elements of the above-described lines / rows or matrices) estimated or determined. For this purpose, the signal chain parameters of this detector 11a or these detectors 11a - 11c so chosen that the entire dynamic range 25 is covered. The signal chain parameters of the following detectors 11b . 11c respectively. 11d ff. are dynamically adjusted based on this measurement so that the SNR for the now limited but well-known dynamic range 26 Optimized is improved.

in the Difference to the second version According to the invention, a TDI calculation takes place here.

This will carry all non-saturated detectors 11 - these are usually all detectors due to the dynamic adaptation 11 except for the selected first for the entire dynamic range 25 - For generating the sum signal and noise reduction at. The signal chain parameters can be used for each of the following detectors 11b ff. be different or exist groups of detectors 11a c, 11d -f, 11g -i etc. with the same signal chain parameters, but the groups differ from one another in at least one signal chain parameter. Similarly, to determine the signal chain parameters multiple detectors 11 be used with the same or different signal chain parameters. Become multiple detectors 11a - 11c as a basis for the definition of the signal chain parameters of the following detectors 11d ff., so here is the possibility ei TDI calculation given, inter alia, to increase the quality of the estimate of the expected signal level.

• 5) In einer fünften Variante der Erfindung sollen die Mischformen der bisher beschriebenen Ausprägungen behandelt werden.

This should be done with the help of 4 be clarified. In the example shown there, three TDI stages were used. With the first low gain TDI stage corresponding to the curve K ₁ , a measurement was taken covering the total dynamic range to be covered 25 for the two subsequent measurements of the following TDI stages could restrict so far that for the then estimated limited dynamic range 26 optimal gain settings could be made. As a result, the SNR is over the entire dynamic range 25 , but especially in the lower and middle regions of the curves K ₃ and K ₄ optimized. In practical implementation, it may well be useful to have more than one detector 11 with low gain factor. This gives the appropriate time to perform the gain settings for the subsequent detector stages (with the SNR curves K ₃ and K ₄ ).

• 5) In a fifth variant of the invention, the mixed forms of the previously described forms are to be treated.

• 6) Eine sechste Ausführungsvariante der Erfindung betrifft die Möglichkeit, Detektoren 11 unterschiedlicher Größe z.B. mit einem Detektor 11a mit viermal so großer Fläche wie die weiteren Detektoren 11b im Sinne der Ausführungen 1) bis 5) auszuwerten.

So it is possible only for a part of the detectors 11 to dynamically vary the signal chain parameters according to the embodiments 2) or 4), and to predetermine the signal chain parameters statically according to design 1) or 3) for the other part. Furthermore, it is possible only a part of the detectors 11 , preferably the detectors 11 with limited dynamic range 26 to use for a TDI calculation according to embodiment 3) or 4) and the other part, preferably for high signal strengths or for the auxiliary measurement for setting the corresponding signal chain parameters of the aforementioned detectors 11 to operate in the sense of execution 1) or 2).

• 6) A sixth embodiment of the invention relates to the possibility of detectors 11 different size eg with a detector 11a with four times as large area as the other detectors 11b in the sense of embodiments 1) to 5).

Example:

An instrument should be designed to record the surface temperatures of areas on planets. Surface temperatures of both water and land areas are to be recorded. Over land areas the possibility must be planned to both observe fire and to determine the soil temperature. The dynamic range to be covered 25 is very large here because the energy emitted by the surface areas in the form of electromagnetic waves (depending on the spectral range under consideration) can vary by more than three orders of magnitude between cold sea surfaces and active sources of fire. At the same time, the applications used here have different requirements for the local resolution. While spatially roughly resolved data can be sufficient across sea surfaces, firefighting usually requires spatially higher resolution image data. These requirements can be met, for example, with the in 5 shown arrangement of detectors 11a and 11b be covered. In 5 are analogous to 3 the direction of movement of the instrument, the scanning direction and the TDI direction entered. Here are the smaller detectors 11a usable for imaging fires while the larger detector element 11b used to observe sea surfaces. Despite the size differences of the individual detector elements can be a linkage of the detectors 11a and 11b according to execution 1) to 5) make sense.

For example, with the help of detectors 11a the signal chain parameters of the detector 11b be set and / or all detectors 11a and 11b contribute in the sense of execution 3) to a common TDI calculation.

• 7). Eine siebte Ausführung der Erfindung beschreibt die Möglichkeit, Anordnungen von Detektoren 11 zu betreiben und im Sinne der Ausführungen 1) bis 5) auszuwerten, bei denen die Detektoren 11a und 11b um Bruchteile des Pixelabstandes (vorzugsweise ½ Pixelabstand) versetzt zueinander ausgelesen werden und dadurch nachträglich eine Erhöhung der Auflösung im abgeleiteten Bild erreicht wird.

The version 6) can additionally be combined freely with the following versions 7) and 8).

• 7). A seventh embodiment of the invention describes the possibility of arrangements of detectors 11 operate and in the sense of the embodiments 1) to 5) evaluate, in which the detectors 11a and 11b by fractions of the pixel spacing (preferably ½ pixel pitch) offset from each other are read out and thereby subsequently an increase in the resolution in the derived image is achieved.

Example:

An imaging instrument requires a high spatial resolution of the resulting image data. At the same time the expected dynamic range is large, but it can also be very small signal levels auftre For radiometric or technological reasons, it would not be possible to reduce the size of the detector surfaces. In this case, there is according to 6 (in combination with prior art subpixel scanning methods) the possibility of having two rows of detectors 11a and 11b offset from each other to read.

In 6 are analogous to 3 the direction of movement of the instrument, the scanning direction (solid arrows) and possible TDI directions (dashed arrows) applied. With the help of the data of these two detector series 11a and 11b As a result, image data is generated which has a somewhat higher spatial resolution than that provided by the individual detectors 11 would be given.

If these staggered rows of detectors are repeatedly executed, then their detectors can be 11c and 11d in terms of characteristics 1) to 5). Furthermore one wins with the detectors 11c already a sufficiently accurate estimate of the expected signal levels of the detectors 11d , This allows the detectors 11c and 11d evaluate each other in the sense of the characteristics 1) to 5).

Be staggered individual rows of detectors 11e and 11f repeatedly executed, can also be evaluated in the sense of the embodiments 1) to 5).

Furthermore one wins with the detectors 11e already a sufficiently accurate estimate of the expected signal levels of the detectors 11f , This allows the detectors 11e and 11f evaluate each other in the sense of the characteristics 1) to 5).

• 8) Eine achte Ausführung der Erfindung beschreibt die Möglichkeit, Detektoren 11 durch vorangestellte unterschiedliche spektrale Filter im Sinne der Ausprägungen 1) bis 5) auszuwerten, indem den aktiven Detektorflächen verschiedene Filter (z.B. Graufilter oder Farbfilter) entweder als fest angebracht oder (z.B. in einem Filterrad) wechselbar installiert sind.

Of course, in this case, repeated read lines of the detectors can be used 11a and 11b . 11c and 11d such as 11e and 11f also evaluate in the sense of the point 2 of the prior art, to lead by means of known methods to a resolution improvement. Furthermore, the versions 6) to 8) are freely combinable with each other.

• 8) An eighth embodiment of the invention describes the possibility of detectors 11 preceded by different spectral filters in terms of the characteristics 1) to 5) by the active detector surfaces different filters (eg neutral density filters or color filters) either as fixed or installed (eg in a filter wheel) changeable installed.

Example:

An optical instrument with a sensor matrix (such as in 3d - 3g indicated for three rows and three columns), which in this case consist of four rows and n columns.

In front Let the first line be a filter with transmission in the range between 400 nm and 500 nm attached, before the second line such with transmission in the range between 500 nm and 600 nm, before the third Line such with transmission in the range between 600 nm and 700 nm and before the fourth line such with transmission between 400 nm and 700 nm. The columns thus represent the scanning width. The Represent rows four spectral channels the fourth being a panchromatic channel.

The detectors 11 The first to fourth lines can be meaningfully linked after execution 1) to 5). Advantageously, detectors 11 the first to third line in the sense of embodiment 3) of the invention with detectors 11 the fourth line contribute to a common TDI calculation. In this case, this would lead to an improvement of the SNR for a panchromatic channel.

The versions 6) to 8) be freely combined with each other.

For a special additional function of the invention is on 7 Referenced.

Used with moving instrument (eg moving camera both in direction of movement and in scan direction) with the detectors 11 a target object 3 Imaged, it comes to smearing, which is also referred to as motion blur. The 7a shows the signal intensity of a target object 3 originating and from a pixel (pixel, detector 11 ) detected radiation 2 while moving over the target object 3 ,

As you can see, the energy of the target object 3 smeared by the movement over a larger area. The standard in optics / optronics degree of distortion and smearing is the Mo modulation transfer function (MTF). The strength of the motion-induced smear and the magnitude of the concomitant reduction in the MTF depends on the size of the integration time. Shortening the integration time will reduce this smearing.

The aim of the previously described embodiments of the invention was to improve the SNR in each relevant dynamic range 25 respectively. 26 , However, if the SNR is improved more than necessary, the resulting reserve can be used to statically or dynamically reduce the integration times and thus to improve the MTF in a targeted manner. This is in 7b outlined.

By shortening the original integration time t ₁ to a shortened integration time t ₂ , the integrated signal levels, in 7a and 7b illustrated as the area under the respective signal intensity curves, reduced in size. Thus, therefore, the SNR deteriorates. As in 7b however, on the other hand, in the case of the shortened integration time t ₂ , the locally "blurred" signal components decrease, and the MTF can thus be improved as follows.

If a specific minimum SNR is defined for the current signal level and the achievable SNR is significantly higher than this, the SNR is reduced to a shortened integration time t ₂ by a defined shortening of the integration time t ₁ , the latter being dimensioned such that the predetermined minimum -SNR (Schellwert 28 ) is still reached or exceeded. This systematically improves the MTF at the expense of the SNR.

in the Following are still advantageous implementation options for the Measuring method according to the invention to be discribed.

Example 1:

Especially in the thermal IR range of the electromagnetic spectrum can be extreme Signal level fluctuations (e.g., local heat sources such as forest fires) become. Usually intended for this a saturation the signal at the output of the instrument can be avoided without the Instrument properties for low signal levels (e.g., normal ambient temperature).

Here, the invention described above can be favorably used by the gain of a TDI stage (detector array) for the fire observation (definition of the entire dynamic range 25 ) and the other (downstream) detectors (other TDI stages) for the observation of relatively cold surfaces such as sea surfaces due to limited dynamic ranges 26 are set up.

As in 8th 2, the SNR in this example is significantly improved in the low to medium signal level range by using Embodiment 3) of the invention by setting the signal chain parameters corresponding to the curves K ₃ and K ₄ in the downstream detectors. By choosing suitable amplification factors, the SNR function in this example is freely adjustable within certain limits. As a result of the combination with the TDI method, the signal levels in the curve areas K ₃ and K ₄ are applied using a gain-dependent weighting (see 10 as hardware embodiment 3) of the invention) links all available signal components. In the process K ₃ signals from detectors 11 used with low and medium gain (detector signals with high gain - because of saturation limit saturation 24 - eliminated) and in the area K ₄ signals of all detectors 11 included so that the SNR increases significantly. In the range of high signal levels (curve K ₁ ) only the signals of the detectors 11 (unsaturated signal processing chains 1 ) processed with low gain.

Example 2:

If the TDI combined measured value calculation is to be performed in hardware, this can be done, for example, in a measured value calculation unit 4 according to the in 10 shown block diagram can be realized. The illustrated hardware concept processes output signals 21 of signal preprocessing chains 1 of the detectors 11 (eg, pixels of a sensor array) according to embodiment 3) of the invention.

The signals of a number n (running index i = 1 ... n) of detectors 11 be first in saturation comparators 41 ₁ to 41 , then check if they are saturated, ie if they are the saturation limit 24 their dynamic range 25 respectively. 26 exceed. Then only the non-saturated signals by means of fixed weights (1 ... n) in the associated multipliers 43 ₁ to 43 _n-1 multiplied. The adders 44 ₁ to 44 _n-1 take all of a target object 3 originating (already weighted) signals together and thus provide the final result for the target object under consideration 3 , Multiplication and addition can be either analog or digital.

The resulting pipeline must exactly match the movement of the individual detectors used for TDI calculation 11 about one and the same target object 3 be adjusted. This boundary condition is in 10 as from the detectors 11 supplied pixel clock r drawn.

Example 3:

According to the embodiment described above 4 ) of the invention, some signal chain parameters and the associated weights can be dynamically adjusted. Is in 11 in a sketched block diagram for a hardware implementation of a measured value calculation unit 4 realized.

Here was the flow chart of 10 extended so that the signal level of the first detector 11 (Signal from pixel 1) by means of a dedicated circuit 45 the amplifiers 46 the following detectors 11 (Pixels 2 to n) controls. Again, a pipeline structure must be introduced, which - like the summation - runs with the pixel clock r. That's what the registers are for 47 intended. The set gain must be from the respective multiplier 43 be taken into account. Therefore, the set gain becomes the multiplier 43 provided as input parameter.

Multiplication and addition can be done both analogue and digital. For digital calculations, the ADC 14 the signal preprocessing chain 1 in the amplifier 46 be integrated.

Example 4:

One possible software-based implementation of embodiment 3) of the invention is in 12 shown. The program flow diagram shown there starts with the initialization of the required memory and weight matrices. After reading in the signals of all detectors 11 In a signal matrix, the elements of the signal matrix are multiplied by the corresponding elements of the weight matrix and stored in a weighted signal matrix. This weighted signal matrix is added element by element to the existing memory matrix. Finally, the last (mth) line of the memory matrix is output. This output represents the result (as a weighted TDI calculation) in the sense of embodiment 3) of the invention.

To the output of the last line of the memory array will be all others Lines of the memory matrix shifted one line down or re-stored. After that, the first line of the memory matrix is zeros initialized to return for the next Read-in process available to stand and to process the following signals in the same way.

1

Signal processing chain

11

detector

12

integrator

13

amplifier

14

Lying to digital converter (ADC)

2

incident radiation

21

output

22

external disorder

23

internal Noise intermediate signal

24

saturation point (the signal level)

25

whole dynamic range

26

limited dynamic range

27

minimal signal level

28

threshold (of the SNR)

3

target

4

Measured value calculation unit

41

Sättigungskomparator

42

SNR comparator

43

multipliers

44

adder

45

circuit (to determine the necessary Gain)

46

gain Control

47

register

K ₁ , K ₂

curves (of the SNR for different parameter combinations)

K ₃ , K ₄

curve areas (of the SNR for Sum of several signal chain combinations)

i

counting index

m

counting index (for not saturated Detector signals)

n

detector number (Pixel number)

r

pixel clock

S _MAX

maximum expected signal level

SNR _MIN

necessary SNR

t ₁

(Original) integration time

t ₂

shortened integration time

Claims (17)

A method for measuring electromagnetic radiation in aerospace instruments for observing the earth or other celestial bodies with a plurality of detectors, each detector with integrator, amplifier and analog-to-digital converter (ADC) forming a signal processing chain, characterized in that parameters of the signal processing chains ( 1 ) of the detectors ( 11 ) are set so that at least one detector ( 11a ) to a dynamic range ( 25 ) for a maximum expected signal level (S _MAX ) of the electromagnetic radiation ( 2 ) and at least one further detector ( 11b ) to a limited dynamic range ( 26 ) is adjusted for small and medium signal levels in order to achieve this limited dynamic range ( 26 ) to obtain a higher signal-to-noise ratio (SNR). Method according to claim 1, characterized in that exclusively measured values of detectors ( 11a ; 11b ) in their signal processing chain ( 1 ) a parameter combination was set with which the current signal level of the incident radiation ( 2 ) below a saturation limit ( 24 ) of the set dynamic range ( 25 . 26 ) the signal processing chain ( 1 ) and detectors ( 11b . 11c ), in their signal processing chain ( 1 ) the saturation limit ( 22 ) was exceeded, excluded from the measurement. A method according to claim 2, characterized in that of several approved for the measurement detectors ( 11a ; 11b ; 11c ) that signal is used in its signal processing chain ( 1 ) a parameter combination is set at which the highest SNR has been achieved. A method according to claim 2, characterized in that of several approved for the measurement detectors ( 11a ; 11b ; 11c ) the one in whose signal processing chain ( 1 ) is set a parameter combination in which a required SNR (SNR _MIN ) has been exceeded by means of a shortest possible integration time to achieve an improved modulation transfer function (MTF) in an object image. Method according to claim 1, 2, 3 or 4, characterized in that the detectors ( 11 ) in the signal processing chains ( 1 ) parameter combinations selected suitable for the application, but fixed. Method according to claim 1, 2, 3 or 4, characterized in that of the detectors ( 11 ) in the signal processing chains ( 1 ) associated parameter combinations of at least one of the parameters of at least one signal processing chain ( 1 ) is set dynamically. A method according to claim 2, characterized in that the approved for the measurement detectors ( 11a ; 11b ; 11c ) are operated in a TDI mode and combined by means of a weighted summation to form a measurement signal, wherein the weighting based on the in the signal processing chains ( 1 ) parameter is set. Method according to claim 7, characterized in that the detectors ( 11 ) in the signal processing chains ( 1 ) assigned parameter combinations for the application selectable, but fixed. Method according to claim 7, characterized in that of the detectors ( 11 ) in the signal processing chains ( 1 ) associated parameter combinations of at least one of the parameters of at least one signal processing chain ( 1 ) is set dynamically. Method according to claim 2, characterized in that individual detectors ( 11a ) with fixed parameters of the signal processing chains ( 1 ) and other detectors ( 11b ; 11c ) are operated with dynamically set parameters. Method according to claim 2, characterized in that at least one of the detectors ( 11a ) in its dimensions of its active area from the remaining detectors ( 11b ; 11c ) is different. Method according to claim 2 or 7, characterized in that the detectors ( 11 ) are operated in a subpixel scan regime, wherein increases in resolution are achieved by shifts of fractions of the average pixel pitch. Method according to claim 1, 2 or 7, characterized in that different detectors ( 11a ; 11b ; 11c ) Various filter elements are connected upstream. Method according to claim 7, characterized in that that the weighted summation is implemented in analog hardware accomplished becomes. Method according to claim 7, characterized in that that the weighted summation is implemented in digital hardware accomplished becomes. Method according to claim 7, characterized in that that the weighted summation is implemented in software. A method according to claim 3, characterized in that at an unnecessary high SNR above the SNR threshold ( 28 ) in the associated signal processing chain ( 1 ) the original integration time (t ₁ ) to a shortened integration time (t ₂ ) for which the SNR is still above the SNR threshold ( 28 ), thereby reducing the reading of the next associated detector ( 11 ) to achieve an increase in the MTF of the object image.